U.S. patent number 4,908,578 [Application Number 07/290,075] was granted by the patent office on 1990-03-13 for method of and device for generating interleaved multiple-slice multiple-echo pulse sequences for mri.
This patent grant is currently assigned to U.S. Philips Corporation. Invention is credited to Filips Van Liere.
United States Patent |
4,908,578 |
Van Liere |
March 13, 1990 |
Method of and device for generating interleaved multiple-slice
multiple-echo pulse sequences for MRI
Abstract
MRI utilizes so-called multiple-slice multiple-echo pulse
sequences. Pulse sequences are generated in order to produce echo
resonance signals in different sub-regions of a body, after which
images of the various sub-regions are reconstructed from the
resonance signals. The pulse sequences, for example spin echo
sequences, are successively generated for the various sub-regions.
More than one echo resonance signal can be generated by means of
each pulse sequence. In order to reduce the overall measuring time
the invention proposes the interleaving of a pulse sequence (ex1,
ep11, er11, ep12, er12) for a sub-region with pulse sequences
(ep02, er-12, ex2, ep21, er21, er02, ex3, ep31, er31, ep22) for
other sub-regions. It is essential that the excitation pulse (ex1),
the echo pulses (ep11, ep12) and the echo resonance signals (er11,
er12) are phase coherent within a pulse sequence.
Inventors: |
Van Liere; Filips (Eindhoven,
NL) |
Assignee: |
U.S. Philips Corporation (New
York, NY)
|
Family
ID: |
19851147 |
Appl.
No.: |
07/290,075 |
Filed: |
December 23, 1988 |
Foreign Application Priority Data
|
|
|
|
|
Dec 24, 1987 [NL] |
|
|
8703127 |
|
Current U.S.
Class: |
324/309;
324/307 |
Current CPC
Class: |
G01R
33/4835 (20130101) |
Current International
Class: |
G01R
33/54 (20060101); G01R 033/20 () |
Field of
Search: |
;324/300,307,309,310,311,312,313,314,318,319,322 ;128/653 |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Tokar; Michael J.
Attorney, Agent or Firm: Briody; Thomas A. Haken; Jack E.
Slobod; Jack D.
Claims
What is claimed is:
1. A method of determining a nuclear magnetization distribution
from magnetic resonance signals which are generated in plural
three-dimensional sub-regions in a body by the application, in the
presence of a steady, uniform magnetic field, of cycles of
respective plural sub-region-selective sequences of RF
electromagnetic field pulses and magnetic field gradient waveforms,
in three mutually orthogonal directions, superimposed on said
steady, uniform magnetic field, which respective plural
sub-region-selective sequences are all applied in each cycle, each
sub-region-selective sequence comprising an RF excitation pulse in
the presence of a selection gradient waveform associated with said
sequence, for selection in a first direction, followed by a phase
encoding gradient waveform, which is varied in amplitude, in a
second and/or a third direction, from one cycle to the next for
phase encoding, followed by one or more RF echo pulses which
generate respective one or more echo resonance signals, wherein
first and second of said plural sub-region-selective sequences are
interleaved in time with each other in that the RF excitation pulse
of the second sub-region-selective sequence occurs in the time
interval between the RF excitation pulse and the last of said one
or more echo resonance signals generated in the first
sub-region-selective sequence and wherein within each
sub-region-selective sequence said RF excitation pulse and said one
or more RF echo pulses are phase coherent with each other.
2. A method as claimed in claim 1, wherein the first of said one or
more echo pulses in the second sub-region-selective sequence occurs
in the time interval between the RF excitation pulse in the second
sub-region-selective sequence and the last echo resonance signal
generated in the first sub-region-selective sequence.
3. A method as claimed in claim 1, wherein there are first and
second RF echo pulses in each sub-region-selective sequence
generating respective first and second echo resonance signals, and
wherein the RF excitation pulse, first RF echo pulse and first echo
resonance signal generated in the second sub-region-selective
sequence occur in the time interval between the second RF echo
pulse and the second echo resonance signal generated in the first
sub-region-selective sequence.
4. A method as claimed in claim 2, wherein there is only one RF
echo pulse in each sub-region-selective sequence generating only
one echo resonance signal.
5. A method as claimed in claim 1, wherein one of said one or more
RF echo pulses of a third sub-region-selective sequence occurs in
the time interval between the between the RF excitation pulse and
the last echo resonance signal generated in the first
sub-region-selective sequence.
6. A method as claimed in claim 5, wherein all gradient waveforms
not associated with the third sub-region-selective sequence which
have a dephasing or rephasing effect in the third
sub-region-selective sequence are chosen to together have equal and
opposite effects respectively preceding and following the said one
of the one or more RF echo pulses of the third sub-region-selective
sequence.
7. A method as claimed in claim 2, wherein one of said one or more
RF echo pulses of a third sub-region-selective sequence occurs in
the time interval between the between the RF excitation pulse and
the last echo resonance signal generated in the first
sub-region-selective sequence.
8. A method as claimed in claim 3, wherein one of said one or more
RF echo pulses of a third sub-region-selective sequence occurs in
the time interval between the between the RF excitation pulse and
the last echo resonance signal generated in the first
sub-region-selective sequence.
9. A method as claimed in claim 1 wherein, within a cycle, the
gradient waveforms and the times of application of RF echo pulses
are the same in the successive time intervals between successive RF
excitation pulses.
10. A method as claimed in claim 2 wherein, within a cycle, the
gradient waveforms and the times of application of RF echo pulses
are the same in the successive time intervals between successive RF
excitation pulses.
11. A method as claimed in claim 1 wherein each RF excitation pulse
is preceded by an inversion pulse.
12. A method as claimed in claim 10 wherein each RF excitation
pulse is preceded by an inversion pulse.
13. A method as claimed in claim 1 wherein the phase encoding
gradient waveform is varied in amplitude from one cycle to the next
for phase encoding in the second direction.
14. A method as claimed in claim 1 wherein the phase encoding
gradient waveform is varied in amplitude in either the second or
the third direction from one cycle to the next for phase encoding
in both the second and third directions.
15. A method as claimed in claim 1 wherein each cycle begins or
ends with a plurality of dummy sequences.
16. A device for determining a nuclear magnetization distribution
from magnetic resonance signals to be generated in plural
three-dimensional sub-regions in a body comprising: means for
generating a steady, uniform magnetic field; means for phase
coherent generating of RF electromagnetic field pulses; means for
generating field gradient waveforms, in three mutually orthogonal
directions, superimposed on said steady, uniform magnetic field;
control means for controlling the means for generating the steady
uniform magnetic field and for controlling the means for generating
field gradient waveforms to cause the application in cycles of
respective plural sub-region-selective sequences of RF
electromagnetic field pulses and magnetic field gradient waveforms,
which respective plural sub-region-selective sequences are all
applied in each cycle, each sub-region-selective sequence
comprising an RF excitation pulse in the presence of a selection
gradient waveform associated with said sequence, for selection in a
first direction, followed by a phase encoding gradient waveform,
which is varied in amplitude, in a second and/or a third direction,
from one cycle to the next for phase encoding, followed by one or
more RF echo pulses which generate respective one or more echo
resonance signals, wherein first and second of said plural
sub-region-selective sequences are interleaved in time with each
other in that the RF excitation pulse of the second
sub-region-selective sequence occurs in the time interval between
the RF excitation pulse and the last of said one or more echo
resonance signals generated in the first sub-region-selective
sequence and wherein within each sub-region-selective sequence said
RF excitation pulse and said one or more RF echo pulses are phase
coherent with each other; means for receiving, detecting and
sampling the echo resonance signals; and processing means which
include programmed arithmetic means for determining the nuclear
magnetization distribution from the sampled echo resonance signals.
Description
BACKGROUND OF THE INVENTION
The invention relates to a method of determining a nuclear
magnetization distribution from magnetic resonance signals which
are generated in a body which is situated in a steady, uniform
magnetic field, which magnetic resonance signals are generated in
sub-regions of the body by means of selective pulse sequences, in a
pulse sequence for generating resonance signals in the sub-region
there being excited nuclear spins by application of a selective RF
electromagnetic excitation pulse, after which at least one magnetic
field gradient which is superposed on the uniform magnetic field is
applied, at least one such gradient being variable in amplitude or
direction from one pulse sequence to another, there being applied
an RF electromagnetic echo pulse in order to generate a resonance
signal from the excited nuclear spins, after which the pulse
sequences are repeated a number of times for different values of
the variable magnetic field gradients and subsequently the nuclear
magnetization distribution is determined from the resonance signals
generated.
The invention also relates to a device for determining a nuclear
magnetization distribution from magnetic resonance signals to be
generated in a body, which device comprises means for generating a
steady, uniform magnetic field, means for generating selective RF
electromagnetic pulses, means for generating at least one magnetic
field gradient whose amplitude or direction are variable, and
control means for controlling the means for generating the
selective RF electromagnetic pulses, means for receiving, detecting
and sampling the magnetic resonance signals, and also comprises
processing means which include programmed arithmetic means for
determining the nuclear magnetization distribution from the sampled
resonance signals.
A method and device of this kind are known from U.S. Pat. No.
4,665,367. According to such a method and device, a body to be
examined is arranged in a steady, uniform magnetic field B.sub.0
whose direction coincides with the z-axis of a stationary cartesian
coordinate system (x, y, z). Under the influence of the magnetic
field, a small excess of the nuclear spins present in the body are
directed in the same way with respect to the theoretically possible
saturation value (all nuclear spins) due to thermal movement. From
a macroscopic point of view, the small excess is to be considered
as a magnetization M of the body or as a slight polarization of the
nuclear spins. After the body arranged in the magnetic field has
been irradiated by an RF electromagnetic pulse which must have a
given frequency, the equilibrium of the magnetization M is
disturbed so that it starts to perform a precessional motion about
the magnetic field B.sub.0. When the processional motion is
observed from a cartesian coordinate system (x', y', z') which
rotates in the same direction and whose z'-axis coincides with the
z-axis of said stationary cartesian coordinate system and when the
angular velocity of the cartesian coordinate system rotating in the
same direction is chosen to be equal to the angular frequency
.omega. of the RF electromagnetic pulse, the magnetization M is to
be considered to be a vector when the angular frequency .omega. of
the RF electromagnetic pulse equals the resonance frequency
.omega..sub.0 of the nuclear spins, which vector moves under the
influence of the irradiation in a plane perpendicular to the
direction of irradiation. The component of the magnetization M
perpendicular to the z-axis, the so-called transverse
magnetization, causes a resonance signal after irradiation. For the
resonance frequency .omega..sub.0, the so-called Larmor equation
.omega..sub.0 =gamma.B.sub.0, holds good, where gamma is the
gyromagnetic ratio of, for example, protons. The angle of rotation
of the magnetization M, and hence the magnitude of the resonance
signal, is determined by the area underneath the RF electromagnetic
pulse. An RF electromagnetic pulse which rotates the magnetization
M through 90.degree. in the stationary coordinate system will be
referred to hereinafter as a 90.degree. pulse. After irradiation,
the magnetization M will relax with a time constant T.sub.1, the
so-called longitudinal relaxation time, until is reaches the static
of equilibrium. A further time constant is the so-called transverse
relaxation time T.sub.2, which is the time constant indicating the
decay of the transverse magnetization. In practical cases the
transverse magnetization decays with a time constant T.sub.2 *
which is substantially smaller than T.sub.2 due to dephasing under
the influence of inevitably present field inhomogeneities. However,
within the relaxation with T.sub.2 always resonance signals can be
obtained by rephasing. By application of magnetic field gradients
G.sub.x, G.sub.y and G.sub.z on the magnetic field B.sub.0, the
field directions of said gradients corresponding to that of the
magnetic field B.sub.0 and their gradient directions extending
perpendicularly to one another, a location-dependent magnetic field
B=B.sub.0 +G.sub.x.x+G.sub.y.y+G.sub.z.z can be generated. U.S.
Pat. No. 4,665,367 describes how resonance signals can be generated
in sub-regions of the body by means of selective pulse sequences.
Selective pulse sequences are pulse sequences in which excitation
pulses occur which excite only the nuclear spins of a sub-region in
the presence of a gradient and which do not excite the nuclear
spins of other sub-regions. The excitation pulses then cover a
range of Larmor frequencies associated with a local field. The
gradient which provides selection, together with the RF
electromagnetic pulses, is also referred to as the selection
gradient (for example G.sub.z). FIGS. 3 and 10 of said U.S. Pat.
No. 4,665,367 show selective excitation for sub-regions
(multiple-slice) and appropriate pulse sequences, respectively. For
16 sub-regions resonance signals are collected while varying the
amplitude of one or two magnetic field gradients (for 2D and 3D
imaging, respectively). In the present example, 4 resonance signals
can be generated in each waiting period required for relaxation of
the magnetization to the state of equilibrium. In that case, 4
waiting periods are required for generating a resonance signal for
all 16 sub-regions. By repeating the sequence a number of times
(for example, 256 times) while varying Gy, a sufficient number of
resonance signals can be collected for determining, for example for
each sub-region, a nuclear magnetization distribution from sampling
values of the respective resonance signals by means of, for example
2D Fourier transformation (Fourier zeugmatography). By repeating
the sequence also while varying G.sub.z, 3D Fourier transformation
can be applied in order to determine a 3D nuclear magnetization
distribution of the body. It will usually be desirable to form
nuclear magnetization distributions of a number of sub-regions (for
example, slices) of the body, and also images thereof in which a
T.sub.1 weighting operation (T.sub.1 contrast) and a T.sub.2
weighting operation (T.sub.2 contrast) are performed. For example,
different echo times must then be used for T2 weighting and
different pulse sequence repetition times and/or inversion pulses
must be used for T.sub.1 weighting. The echo time is the period of
time expiring between the generating of the excitation pulse and
the occurrence of an echo resonance signal in a sequence. In the
case of a comparatively long echo time, for example as in the case
of T.sub.2 weighting, a loss of time is incurred. T.sub.1 weighting
and T.sub.2 weighting together offer a suitable discrimination of
tissue in, for example, in vivo measurements.
SUMMARY OF THE INVENTION
It is an object of the invention to provide a method whereby a
substantial reduction of the measuring time is achieved.
A method in accordance with the invention is characterized in that
RF electromagnetic pulses and magnetic resonance signals in a pulse
sequence for generating resonance signals in a sub-region are
interleaved in time with RF electromagnetic pulses and magnetic
resonance signals in pulse sequences for generating magnetic
resonance signals in at least one other sub-region, it being
ensured that the excitation pulse and the echo pulses are phase
coherent within a pulse sequence. In the case of comparatively long
echo times, time intervals which are not necessary for the
switching of gradients of a pulse sequence associated with a
sub-region are used for generating pulses and gradients of other
sub-regions. The method as described in U.S. Pat. No. 4,665,367 can
be extended in known manner so as to obtain a method whereby
T.sub.1 -weighted as well T.sub.2 -weighted images can be made by
generating, after the occurrence of the first echo resonance
signal, a second echo pulse for generating a second echo resonance
signal (multiple-echo) and by reconstructing images from groups of
resonance signals generated by means of the first echo pulses and
the second echo pulses, respectively. When more than one echo pulse
is applied in this manner in order to obtain T.sub.1 -weighted and
T.sub.2 -weighted images, mainly the intervals succeeding the echo
pulses are used. The first echo pulse will then generally succeed
the excitation pulse so closely that the interval between the
excitation pulse and the first echo pulse of a sequence cannot be
used for other sequences. The time required for transverse
relaxation will then be used for the interleaving of pulse
sequences associated with different sub-regions. By chosing all
corresponding echo times of the sub-regions to be equal in the
method in accordance with the invention, comparable images are
obtained after reconstruction.
It is to be noted that multiple-slice and multiple-echo are
separately described in "Principles of MR Imaging", a publication
by Philips Medical Systems, November 1984.
Multiple-slice--multiple-echo is also described in brief in
"Multiple Spin-Echo Imaging with a 2D Fourier Method", by R.
Graumann et al, Magnetic Resonance in Medicine 3, pp. 707-721,
notably on page 716. It is also to be noted that, for example in
U.S. Pat. No. 4,577,152, FIG. 4, first the inversion pulses are
given for different sub-regions in so-called inversion recovery
measurements, and subsequently the resonance signals are
successively generated for the sub-regions by means of 90.degree.
pulses. However, no phase coherence is then required between the
inversion pulse on the one side and the 90.degree. pulse and the
resonance signal on the other side, because the inversion pulse
only inverts the magnetization M and does not render it transverse.
Therefore, this method does not concern interleaved sequences in
the sense of the invention and the resonance signals are
consecutively generated.
A preferred version of a method in accordance with the invention is
characterized in that, in order to obtain suitable rephasing at
echo instants, the gradient waveforms are the same for all
sub-regions, the magnetic resonance signals being obtained by the
phase-coherent variation of the frequency of the excitation pulse
and of the echo pulses in the pulse sequences of the respective
sub-regions. When the gradient G.sub.x applied during the
measurement and sampling of the resonance signal also has the same
sign and the same form for all resonance signals in a sequence, the
conditions in which the various resonance signals are obtained from
a sub-region are constant and hence comparable.
A version of a method in accordance with the invention is
characterized in that in the pulse sequences the excitation pulse
is preceded by an inversion pulse. An inversion recovery pulse
sequence is thus obtained.
Further versions of methods in accordance with the invention are
characterized in that either for the phase encoding of the nuclear
spins in a first direction the amplitude of a first gradient is
varied during the repetition of the pulse sequences, or for the
phase encoding of the nuclear spins in the first and a second
direction the amplitude of the first and a second gradient is
varied, the amplitude of one of said first and second gradients
being varied per pulse sequence, or the amplitude of the first and
a third gradient is simultaneously varied during the repetition of
the pulse sequences, or in the pulse sequences the amplitude of the
first and the third gradient is simultaneously varied during the
repetition of the pulse sequences; this is subsequently repeated
while varying the amplitude of the second gradient. Thus, methods
are achieved with either 2D Fourier transformation or 3D Fourier
transformation, or 2D projection reconstruction or 3D projection
reconstruction.
A device in accordance with the invention is characterized in that
the processing means also comprise programmed arithmetic means for
controlling the control means so as to interleave in time RF
electromagnetic pulses and magnetic resonance signals in a pulse
sequence for generating magnetic resonance signals in a sub-region
with RF electromagnetic pulses and magnetic resonance signals in
pulse sequences for generating magnetic resonance signals in at
least one other sub-region, the device comprising a
phase-continuous synthesizer for the phase coherent generating of
RF electromagnetic pulses in order to maintain phase coherence
between an excitation pulse, echo pulses and echo resonance signals
of a pulse sequence associated with a sub-region of the body. The
method in accordance with the invention can be performed by means
of such a device.
BRIEF DESCRIPTION OF THE DRAWING
The invention will be described in detail hereinafter with
reference to a drawing; therein:
FIG. 1 diagrammatically shows an MRI device for executing the
method in accordance with the invention,
FIG. 2 shows a body with a subdivision into sub-regions,
FIG. 3 shows a multiple-slice multiple-echo pulse sequence,
FIG. 4A shows the phase-continuous variation of the frequency of a
sinusoidal signal,
FIG. 4B further illustrates the phase coherence as used in
accordance with the invention,
FIG. 5 shows time-interleaved pulse sequences in accordance with
the invention for two echo resonance signals per pulse sequence,
and
FIG. 6 diagrammatically shows some further time-interleaved pulse
sequences in accordance with the invention.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
FIG. 1 diagrammatically shows an MRI device 1 for performing the
method in accordance with the invention, comprising magnet coils 2
for generating a steady, uniform magnetic field B.sub.0, gradient
magnet coils 3 for generating a magnetic field gradient G.sub.x,
gradient magnet coils 4 for generating a magnetic field gradient
G.sub.y, gradient magnet coils 5 for generating a magnetic field
gradient G.sub.z, and a transmitter/receiver coil 6 for
transmitting RF electromagnetic pulses to the body and for
receiving magnetic resonance signals from the body, respectively.
When the magnet coils 2 are constructed as resistance magnets, they
are powered by a DC power supply source 7. When the magnet coils 2
are constructed as permanent magnets, of course, the DC power
supply source 7 is absent. The magnet coils 2 can also be
constructed as superconducting magnets. During execution of the
method, the body is arranged within the magnet coils 2. The
gradient magnet coils 3, 4 and 5 are powered, via respective power
supply lines 8, 9 and 10, by a controllable power supply source 11
which is controlled by control means 12. The magnetic field
gradients which are superposed on the magnetic field B.sub.0 can be
independently generated. To this end there are provided three
control lines 13, 14 and 15 between the control means 12 and the
controllable power supply source 11. In the embodiment shown, the
spatial arrangement of the gradient coils is such that the field
direction of the magnetic field gradients G.sub.x, G.sub.y and
G.sub.z coincides with the direction of the magnetic field B.sub.0
and that the gradient directions extend perpendicularly to one
another as denoted by three mutually perpendicular axes x, y and z
in FIG. 1. The control means 12 are coupled to processing means 16
via a number of lines. The processing means 16 are coupled to an
analog transmitter section 17 of a transmitter 18 for transmitting
RF electromagnetic pulses and to an analog receiver section 19 of a
receiver 20 for receiving, detecting and sampling magnetic
resonance signals. The analog transmitter section 17 is coupled,
via a line 21, to a directional coupling device 22 which is
coupled, via a line 23, to the analog receiver section 19. The
transmitter/receiver coil 6 is coupled, via a line 24, to the
directional coupling device 22. The transmitter/receiver coil 6 can
alternatively be constructed as a separate transmitter coil and a
separate receiver coil. In that case the directional coupling
device 22 is dispensed with. The analog transmitter section 17
comprises a transmission frequency oscillator 25, a transmission
frequency mixing stage 26 and an RF power amplifier 27. The
transmission frequency mixing stage 26 is coupled, by way of an
input 28, to an output 29 of the transmission frequency oscillator
25 and is also coupled, by way of an input 30, to an output 31 of a
digital transmitter section 32 of the transmitter 18. Furthermore,
via an output 33 the transmission frequency mixing stage 26 is
coupled to an input 34 of the RF power amplifier 27. The digital
transmitter section 32 is included in the processing means 16, but
can alternatively be constructed as a separate digital unit. The
analog transmitter section 17 and at least a part of the digital
transmitter section 32 may also be accommodated in one unit.
However, this is irrelevant for the operation of the device 1. The
digital transmitter section 32 comprises a frequency-continuous
synthesizer 35, a multiplier device 36, a digital-to-analog
converter 37, and a register 38 for the storage of digital
amplitude information. The phase-continuous synthesizer 35 is
coupled, by way of an output 39, to an input 40 of the multiplier
device 36 which is coupled, by way of an output 41, to an input 42
of the digital-to-analog converter 37. The digital-to-analog
converter 37 is coupled, by way of an output 43, to the input 31 of
the digital transmitter section 32. The register 38 is coupled, by
way of an output 44, to an input 45 of the multiplier device 36.
The phase-continuous synthesizer 35 comprises an address generator
46, a read-only memory 47, a register 48 for the storage of digital
frequency information, and a register 49 for the storage of digital
phase information. The address generator 46 is coupled, by way of
an output 50, to an input 51 of the ROM 47. Via an input 52, the
address generator is also coupled to an output 53 of the register
48 and, by way of an input 54, to an output 55 of the register 49.
The ROM 47 is coupled, by way of an output 56, to the output 39 of
the phase-continuous synthesizer 35. The registers 38, 48 and 49
are coupled, by way of respective inputs 57, 58 and 59, to
programmed arithmetic means 60. The programmed arithmetic means are
also coupled to the control means 12, via a line 61, and to a
display screen 62 for the display of an image of the nuclear
magnetization distribution. The analog receiver section 19 is
coupled, by way of an input 63, to the transmission frequency
oscillator 25 and, by way of an output 64, to an input 65 of the
processing means 16. The analog receiver section which is not
described in detail herein comprises a conventional detector
circuit (not shown). For a more detailed description of a
demodulator based on (double) phase-sensitive detection, reference
is made, for example to an article by P. R. Locher "Proton NMR
Tomography", Philips Technical Review, Vol. 41, 1983/84, No. 3, pp.
73-88. It is to be noted that the transmitter/receiver described
therein does not comprise a phase-continuous synthesizer. The
processing means 16 also comprise at least one analog-to-digital
converter. FIG. 1 shows an analog-to-digital converter 66 whose
input is formed by the input 65 of the processing means and which
is coupled, by way of an output 67, to the programmed arithmetic
means 60. The analog-to-digital converter 66 is controlled by the
control means 12, via a control line 68. For a more detailed
description of a so-called digital transmitter/receiver comprising
a phase-continuous synthesizer, reference is made to the commonly
owned allowed U.S. patent application Ser. No. 196,534, filed May
19, 1988 prepublished Netherlands Patent Application No. . . . (PHN
12.134). European Patent Application EP 0 165 057 also describes
such a digital transmitter/receiver. The method in accordance with
the invention can be performed only if the device 1 comprises a
phase-continuous synthesizer; however, the detection may be
phase-sensitive or phase-insensitive. The generating of a resonance
signal in the body by means of an RF electromagnetic pulse and the
described MRI device will now described. The ROM 47 contains a
sinusoidal function which is stored in digital form in a table at
successive memory locations. The table contains, for example 1024
values of one period of the sinusoidal signal. When the ROM is
passed through cyclically at a uniform speed, a periodic sinusoidal
signal appears on the output 56 thereof. The address generator 46
generates addresses for the ROM 47. At successive addresses (and
with a constant clocking frequency (clock means for the clocking
out of table values are not shown), the minimum frequency is
reached. When a higher frequency is desirable, the ROM memory must
be passed through in larger steps (for example, each time 1, 2, 3,
. . . addresses are skipped). It will be apparent that the steps
may not be too large, because an analog sinusoidal signal must be
reconstructed by filtering from the table values obtained. The
known Shannon theorem must then be satisfied. The contents of the
register 48 are decisive as regards the frequency of the sinusoidal
signal (jumps are feasible in the table). Because the programmed
means 60 can load registers 48 and 49 so that the address generator
46 can form the exact address at any instant, phase coherence can
be simply maintained when the frequency is varied. In this respect
reference is also made to a phase-continuous synthesizer known as
"Wavetek", model 5155A. The register 38 contains amplitude
information, which means in this case that the digital number in
the register 38 is a measure for the amplitude of the sinusoidal
signal present on the output of the multiplier device 36. When the
control means 12 enable, for example the address generator 46 via
an enable line 69, the output of the digital-to-analog converter 37
will carry a sinusoidal signal whose amplitude, frequency and phase
are determined by the contents of the registers 38, 48 and 49. By
continuously varying the contents of the register 38, the amplitude
of the sinusoidal signal can be modulated in order to generate
pulses and a given bandwidth can be imparted to the pulses. For
example, frequencies between 100 and 700 kHz are generated. In the
transmission frequency mixing stage 26, the (modulated) sinusoidal
signal on the output 43 of the digital-to-analog converter 37 is
mixed with the signal of the transmission frequency oscillator 25
(this oscillator may be, for example a PLL (phase locked loop)
oscillator). The output 33 of the transmission frequency mixing
stage 26 will carry a pulse having such a frequency contents that
it is suitable for exciting nuclear spins present in a steady
magnetic field (for example, when the body contains protons, a
proton resonance signal can be generated; for example, when the
field strength of the magnetic field B.sub.0 amounts to 1.5 T, the
resonance frequency of protons amounts to 63.86 MHz). When a signal
of, for example 63.56 MHz is generated in the transmission
frequency oscillator 25, proton resonance will be obtained on the
output 53 for a signal of 300 kHz when use is made of a 1.5 T MRI
system. By generating a signal having a given bandwidth with
respect to an iso centre in the presence of a magnetic field
gradient G.sub.z superposed on the magnet field B.sub.0, a
sub-region (for example, a slice) of the body can be selectively
excited by means of a selective pulse sequence. The isocentre is
the point within the magnet coils 2 in which exactly the field
strength B.sub.0 occurs when all magnetic field gradients are
activated (in the case of ideal magnet coils 2). The generated
pulse is amplified by means of the power amplifier 27 and is
applied, via the directional coupling device 20, to the
transmitter/receiver coil 6, so that proton resonance signals are
generated in the present example. While varying magnetic field
gradients, a large number of resonance signals are generated. The
generated resonance signals are received, detected and sampled in
known manner and from the sampled signals a nuclear magnetization
distribution is determined in known manner, for example by Fourier
transformation. Subsequently, an image of the nuclear magnetization
distribution is displayed on the display screen 62 by converting,
for example signal values into grey tones.
FIG. 2 shows a body 1 with a sub-division into sub-regions, for
example "slices" d1 to d16.
In FIG. 3 pulses and echo resonance signals of known multiple-slice
multiple-echo pulse sequences for the selective generating of echo
resonance signals in the slices d1 to d16 are plotted as a function
of time T. The associated gradients are not shown. FIG. 5 of the
U.S. Pat. No. 4,665,367 shows so-called spin echo pulse sequences
in which field gradients are also shown for 2D Fourier
zeugmatography. The description of FIG. 3 of the present
application will now be continued. Using the pulse sequence shown,
nuclear spins of other slices can be excited within a waiting
period TR for the return to a state of equilibrium of the
magnetization M of a slice after the excitation thereof. First, for
example slice d1 is excited by means of a selective excitation
pulse ex1 (the pulse is given in the presence of a selection
gradient G.sub.z (not shown)). Subsequently, an echo pulse ep11 is
applied in order to generate a first echo resonance signal as by
means of the spin echo pulse sequence described in the U.S. Pat.
No. 4,665,367. The excitation pulse ex1 is a 90.degree. pulse and
the echo pulse is a 180.degree. pulse, which means that the
magnetization M is rotated through 90.degree. and 180.degree.,
respectively. An echo resonance signal er11 occurs one echo time
TE1 after the application of the excitation pulse ex1. The nuclear
spins of the slice d1 then continuously dephase due to the field
inhomogeneities of the magnetic field B.sub.0. By rephasing the
nuclear spins with an echo pulse ep12, a second echo resonance
signal er12 is generated, which echo resonance signal occurs one
echo time TE2 after the excitation pulse ex1. T2 relaxation is
denoted by a broken line. For as long as transverse magnetization
still exists, echo resonance signals can be generated by means of
echo pulses. The echo resonance signals themselves relax with the
previously said relaxation time T.sub.2 * due to the inevitable
field inhomogeneities. Depending on the waiting period TR, echo
resonance signals can be generated for other slices. Within the
waiting period TR, for example 4 slices are excited. ex2 is a
(selective) excitation pulse for slice d2, ep21 is a first echo
pulse for slice d2. o indicates an interruption of the time axis.
The references ex4, ep41, ep42, er41 and er42 denote the excitation
pulse, echo pulses and echo resonance signals, respectively, for
the slice d4. In order to obtain two resonance signals for each
slice, 4 waiting periods are required. By repeating the described
pulse sequences for different values of a phase encoding gradient
G.sub.y (for example, 256 times) and by sampling the resonance
signals obtained (for example, 256 samples per resonance signal),
for example 2.times.16 2D images of the 16 slices can be obtained
after grouping and Fourier transformation of the sampling values,
which images exhibit mainly T1 contrast and mainly T2 contrast,
respectively. The slices need not be successively measured. From a
point of view of interference, it is advantageous to measure the
slices in a staggered order, for example first the slices d1, d5,
d9 and d13 and subsequently the slices d2, d6, d10 and d14, and so
on.
FIG. 4A shows the phase-continuous variation of the frequency of a
sinusoidal signal as a function of the time t. The phase-continuous
synthesizer 35 generates a sinusoidal signal having a frequency
.omega.1 at the instant t0. At the instant t1, the frequency
becomes .omega.2. At the instant t2 the frequency is reset to
.omega.1. The interrupted line o indicates that the phase of the
sinusoidal signal having a frequency .omega.1 is phase coherent,
during the interval from t0 to t1, with the phase of the sinusoidal
signal having the frequency .omega.1 in the interval after t2. The
programmed arithmetic means 60 ensure, in conjunction with the
registers 48 and 49, that phase coherence always exists.
FIG. 4B illustrates phase coherence as used in accordance with the
invention. .phi.=.omega..multidot.t is plotted as a function of the
time t for a number of slices. In accordance with the invention,
pulse sequences of different slices are interleaved in time. This
necessitates phase coherence between excitation pulse, echo pulses
and echo resonance signals within each pulse sequence. At the
instant t0, a pulse sequence for the slice d1 commences. The heavy
line segment ls1 indicates that, for example a 90.degree.
excitation pulse for the slice d1 is supplied. At the instant t1, a
pulse sequence commences for slice d2 with a 90.degree. excitation
pulse, denoted by the line segment ls2. At t2 the pulse sequence
for the slice d1 is continued, indicated by the line segment ls3.
For example, a 180.degree. echo pulse is applied. It is an
essential aspect of the method in accordance with the invention
that phase coherence exists between the 90.degree. excitation pulse
in ls1 and the 180.degree. echo pulse in ls3. If the phase were not
correlated, the Fourier transformation would not produce useful
results, because such a transformation is essentially a correlation
technique.
FIG. 5 shows a fundamental portion of a cycle of time-interleaved
slice-selective pulse sequences in accordance with the invention
for two echo resonance signals slice-selective per pulse sequence
as a function of the time t. In the example shown, pulse sequences
are described for a 2D Fourier zeugmatography. The time t is given
in ms. Plotted as a function of the time t are successively: RF
electromagnetic pulses hf, a selection gradient G.sub.z, a phase
encoding gradient G.sub.y, a measuring gradient G.sub.x, echo
resonance signals MR, data gating signals g, and a time axis ts.
For the description of the pulses the same notation is used as for
the description given with reference to FIG. 3. The pulse sequence
1 spanning from excitation pulse ex1 to the second echo resonance
signal er12, for generating echo resonance signals in a first slice
is fully illustrated. At t=0 a selective RF electromagnetic
excitation pulse ex1 is generated in the transmitter 18. Via the
control line 15, the controllable power supply source 11 is
activated in order to generate the selection gradient G.sub.z
during the pulse ex1. In the transmitter 18 a given bandwidth is
imparted to the pulse ex1 by modulation, via the register 38, of
the sinusoidal signal generated by means of the phase-continuous
synthesizer 35. The bandwidth and the selection gradient G.sub.z
are matched in known manner (see, for example Locher page 83). At
t=10 ms, a first selective echo pulse ep11 is generated in the
transmitter 18. The control line 15 activates G.sub.z again.
Dephased nuclear spins rephase at t=20 ms, at which instant a first
echo resonance signal er11 arises in the body. Between the
excitation pulse ex1 and the echo pulse ep11, G.sub.y is activated,
via the control line 14, for the phase encoding of the nuclear
spins. Via the control means 12, the amplitude of G.sub.y can be
varied. After demodulation in the analog receiver section 19, the
echo resonance signal er11 is sampled, during the data gating g11,
by means of the analog-to-digital converter 66 (during the data
gating g11, for example 256 samples are taken). For frequency
encoding, G.sub.x is activated, via the control line 13, during the
occurrence of the echo resonance signal er11. G.sub.x is also
activated between the pulses ex1 and ep11. At t=80 ms, the
transmitter 18 generates a second selective echo pulse ep12, so
that a second echo resonance signal er12 occurs at t=140 ms. It is
essential that the phase is coherent for the pulse sequence 1, so
the phase must be coherent around t=0, t=10 ms, t=20 ms, t=80 ms
and t=140 ms. The transmission frequency and the demodulation
frequency may differ, but phase coherence is required. The
deviation of the transmission frequency and the receiving frequency
depends on the transmitter/receiver used. When use is made of a
phase-sensitive detector during demodulation, the phase of the
excitation pulse ex1 in the pulse sequence 1 is irrelevant. When a
digital transmitter/receiver is used, the phase of the excitation
pulse is important. All excitation pulses can then have, for
example a phase zero. The gradients switched during the pulse
sequence 1 are G.sub.zex1, G.sub.x1, G.sub.y1, G.sub.zep11,
G.sub.xer11, G.sub.zep12, and G.sub.xer12. In the present example
the echo time of the first echo resonance signal er11 amounts to 20
ms and the echo time of the second echo resonance signal er12
amounts to 140 ms. The time interval between the pulses ex1 and
ep11 is used substantially completely for the switching of
gradients. The time interval between the pulses ep11 and ep12 and
the time interval between the pulse ep12 and the echo resonance
signal er12 are not used for that purpose. The latter time
intervals are used for pulses and/or resonance signals of other
slices. Use is made of the time required for the transverse
relaxation time T.sub.2 between different echo resonance signals.
At t=50 ms, the transmitter 18 generates a selective 90.degree.
excitation pulse ex2 of a pulse sequence 2 for a second slice in
order to generate a resonance signal (FID, not shown); at t=50 ms,
it generates an echo pulse ep21 and at t=130 ms an echo pulse ep22;
at t=70 ms, an echo resonance signal er21, the rephased FID signal,
arises in the second slice. The further signals of the pulse
sequence shown in FIG. 5 are the gradients G.sub.zex2, G.sub.y2,
G.sub.x2, G.sub.zep21, G.sub.xer21, and G.sub.zep22, and the data
gating signal g21. It is essential that for the pulse sequence 2
the phase is coherent, i.e. around t=50 ms, t=60 ms, t=70 ms and
t=130 ms. Because in the time interval between the pulses ep21 and
ep22 of the second slice there is also generated, for example the
pulse ep12 of the first slice having a different frequency contents
(synthesizer 35 generates a different frequency), it is essential
that the synthesizer is a phase-continuous synthesizer. FIG. 5 also
shows the pulses ep02, ex3 and ep31, the gradients G.sub.zep02,
G.sub.xer-12, G.sub.xer02, G.sub.zex3, G.sub.y3, G.sub.x3,
G.sub.zep31 and G.sub.xer31, and the data gating signals g-12, g02
and g31. There are, for example sequences for 16 slices in each
cycle. It will be evident that the pulses, gradients and echo
resonance signals of only a few of the 16 slices are shown. For the
other slices, the pulse sequences, having an identical composition,
simply follow from the described pulse sequences. After two
resonance signals have been generated in all slices, the value of
the phase encoding gradient G.sub.y is changed via the control
means 12. The described pulse sequences are repeated for, for
example 256 values of G.sub.y. In the embodiment described it is
attractive to "expose" nuclear spins to the same gradients during
time intervals of a pulse sequence which are used for other pulse
sequences. This is done in order to avoid the drawbacks of
instrumental deficiencies, such as the difference in the effect of
eddy currents. In the present example, for the pulse sequence 1 the
time intervals used for other pulse sequences are situated on both
sides of the echo pulse ep12. The gradient waveforms of G.sub.x,
G.sub.y and G.sub.z associated with the other pulse sequences have
the same, mirror-image shape with respect to the echo pulse ep12.
Furthermore, the conditions in which the various echo resonance
signals of a slice are obtained are constant, which means that the
measuring gradient G.sub.x has the same sign and the same shape for
all echoes of a slice. The same echo times hold good for for all
slices. In the present example they amount to 20 ms and 140 ms. The
images obtained from the echo resonance signals are thus suitably
comparable. For the selected gradient waveforms rephasing occurs
exactly at the echo time for an echo resonance signal. For all
slices the same sequence of gradient waveforms is chosen. Only the
frequency of the synthesizer and the demodulation frequency differ
for each slice. When all resonance signals of slices have been
generated and processed for a given value of the phase encoding
gradient G.sub.y, the value of the phase encoding gradient changes
so that the rephasing condition is no longer satisfied. In order to
restore the rephasing condition, it is necessary to apply a number
of dummy sequences. This number depends on the degree of
interleaving of the pulse sequences, to be denoted by an
interleaving factor I. In the present example I=1. For P steps of
the phase encoding gradient G.sub.y, I.P dummy sequences are
required. If P=256, 256 dummy sequences are required for I=1. This
gives rise to a somewhat smaller reduction of the measuring time.
In comparison with the known multiple-slice multiple-echo sequence,
the reduction of the total measuring time amounts to a reduction
factor R=2.I+1, so in this case a factor 3. Because of said dummy
sequences, this factor will be slightly smaller in practice. The
increase in the total measuring time with respect to the ideal
reduction is smaller as more slices are measured. In the present
example there is a short echo time of 20 ms. In an image to be
reconstructed from echo resonance signals mainly T.sub.1 -contrast
becomes manifest. Furthermore, there is a long echo time of 140 ms
whereby mainly T.sub.2 -contrast is achieved. It is to be noted
that the interleaving factor I is linked to the difference in echo
time. When the difference in echo time is greater, a larger
interleaving factor can be selected.
FIG. 6 diagrammatically shows some portions of cycles of
time-interleaved slice-selective pulse sequences in accordance with
the invention as a function of the time t. For the sake of clarity,
only the RF electromagnetic pulses and the echo resonance signals
are shown. A short, upwards directed stroke represents a 90.degree.
excitation pulse and a long, upwards directed stroke represents a
180.degree. echo pulse; a short downwards directed stroke
represents an echo resonance signal. Furthermore, the same notation
is used as in FIG. 3. For example, ex1 denotes an excitation pulse,
ep11 denotes a first echo pulse, ep12 denotes a second echo pulse,
er11 denotes a first echo resonance signal and er12 denotes a
second echo resonance signal of a pulse sequence for a first slice.
The first line r1 shows pulse sequences where for the interleaving
factor I=2, so that, ignoring the necessary dummy sequences, a
reduction factor R=5 is obtained. Two echo resonance signals E=2
appear per pulse sequence. On the second line r2, I=1, R=3 and E=2;
on the third line r3, I= 1, R=3 and E=3; on the fourth line r4,
I=1, R=3 and E=1; on the fifth line r5, I=2, R=5 and E=1. The
gradients can be switched in the same way as shown in FIG. 5 for 2D
Fourier pulse sequences. When the sequences are also repeated while
G.sub.z is varied after the 90.degree. excitation pulse in order to
achieve phase encoding of the nuclear spins in a second direction,
3D Fourier pulse sequences are formed. The pulse sequences can also
be made suitable for 2D and 3D projection reconstruction. G.sub.x
and G.sub.y are then simultaneously varied for 2D projection
reconstruction after the excitation pulse ex. For 3D projection
reconstruction these pulse sequences are also repeated while
varying G.sub.z after the 90.degree. excitation pulse. 2D
projection reconstruction is described in detail in the cited
article by Locher. 3D projection reconstruction is a method derived
therefrom. The pulse sequences can be rendered suitable for
inversion recovery measurements, being measurements for obtaining
information concerning the relaxation time T.sub.1. In that case
the excitation pulse in each pulse sequence should be preceded by a
180.degree. inversion pulse. It is to be noted that no phase
coherence is required between the 180.degree. inversion pulse and
the remainder of the pulse sequence, because the 180.degree.
inversion pulse does not cause transverse magnetization so that no
dephasing/rephasing conditions need be satisfied.
Within the scope of the invention many alternatives will be
apparent to those skilled in the art. In order to ensure that
within a pulse sequence phase coherence is achieved within a pulse
sequence for the time-interleaved pulse sequences, for example use
can also be made of more than one synthesizer 35, i.e. one
synthesizer for each sub-region to be selected. In that case such a
synthesizer need not be phase continuous. The method in accordance
with the invention is also carried out by activating, using the
control means 12, a multiplex switch connected between the
synthesizers 35 and the multiplier device 36 during the generating
of the pulse sequences so that the associated synthesizer is
activated for each sub-region. When the MRI device comprises such a
multiple-synthesizer (for example, 16 synthesizers), the
synthesizers may have a digital as well as a conventional analog
construction. Such a solution, however, involves substantially more
hardware so that this solution is not to be preferred.
* * * * *